12 Stages of Big Data Grief

There are many obstacles to turning data into actionable insight, but you can overcome many of them by starting from a rich and accurate source.

Most organizations are some way along a digital transformation journey that is changing how they run their operations as well as the way they interact with customers and partners. As more transactions depend on optimized infrastructure that can deliver the consistency and performance a modern business demands, the way digital assets are managed becomes business critical.

Organizations realize that the game has shifted on what’s important and they require new strategies, assets, processes, and KPIs to effectively assess, manage, and grow their digital business. Demographic targeting has become much more narrow and precise. Measuring delivery cycles in months and years is no longer acceptable. Cyber security is a core business discipline. Customer interaction must be immediate, continuous, and on demand or customers change providers. Employees expect more and customers demand more.

Technology infrastructure, once the substantial portion of the iceberg below the surface, not to be seen or understood by the business, is highly relevant to business leaders. Managing a digital business requires a new kind of intelligence – digital intelligence – that allows business leaders to correlate aspects of the commercial and business activities with customer experience and outcomes with underlying technology performance and security.

As with most things nowadays, the answers lie in data – or at least in the analysis of data. The good news is that the process of digital transformation generates huge amounts of data. Organizations know that this data is gold for the multiple stakeholders tasked with optimizing strategy, operations, performance, customer management and maintaining security. The bad news is that they are struggling to manage it let alone extract insight from it.

There is a litany of challenges that we call the 12 Stages of Big Data Grief:

  1. Identify the right data sources. It starts with identifying the questions you want to answer but then you have to find and source the data that can help provide the answers.
  2. Clean, normalize and prepare data. If you can’t ensure the accuracy, reliability, consistency and context of your data, you could be heading for project failure. By the way, when an unknowing application or database administrator changes the source format, expect to re-normalize.
  3. Select appropriate data analysis technologies. Big Data file formats alone, like Orc and Hive, offer bewildering choices that demand specialist knowledge. Questions around types of data, types of interactions, volumes, calculations, etc. are all relevant here as well as planning an overall schema that will likely change a dozen or so times.
  4. Build or integrate data analysis infrastructure. The expense is non-trivial and implementation is difficult. Many organizations lack deep data analytics technology skill sets and don’t have time to learn the critical lessons firsthand and then rearchitect.
  5. ETL (Extract, Transform and Load) complexity and integration issues. A common complaint is that by the time people get their hands on the data it’s already out of date.
  6. Initial analysis, data integrity and ongoing reliability checks. The need to constantly review and ‘test’ results is a continuous requirement that sucks up time and resources.
  7. Find and integrate better quality data. If outcomes from Stage 6 are poor, it’s back to the beginning – identifying, collecting and collating new data sets.
  8. Deal with scope creep. A risk along every step of a Big Data project is that budgets are exceeded and time is wasted as people cram more and more into the panacea of the big data project. So build for a purpose and stick to it. Think in terms of resulting workflows from the start.
  9. Manage the data analytics infrastructure roadmap. It‘s hard to plan future investments when the technology is evolving and your business needs are changing.
  10. Maintain results quality with an evolving data management roadmap. Expertise and innovation will be in constant demand if analytics are to stay relevant. If they’re not relevant, they’re redundant.
  11. Advanced analytics with AI and machine learning. This requires data science expertise that’s expensive to recruit and maintain, and may be limited in their effectiveness without domain knowledge
  12. Visualization and reporting. Data identification, collection and collation are only part of the challenge. Unless you can surface the results and intuitively reveal insights that can be actioned, all the hard work is wasted.

None of these stages are easy, but there’s no doubt that heavy lifting occurs at the start when data is initially sourced. When it comes to analytics, the quality, precision and normalization of the data is a foundational building block (garbage in, really expensive garbage out).

Sometimes, the best sources of data are unexpected ones. At Corvil, we passively capture and extract information from data in motion over the network – at scale and in real-time. Granular detail reveals not just that a transaction occurred but the precise time, how long it took and, often, even the customer or user.

We precision timestamp and sequence that information, normalize it, enrich and analyze it, making it available to stream in real-time to business, security and IT solutions as well as big data analytics platforms. And, because the data is voluminous, contextual, and normalized, it provides a strong foundation for applying machine learning techniques.

With Corvil Intelligence Hub, the grief is gone.

A rapidly deployable, self-service solution replaces the effort of securing the triumvirate skills of data analytics technology, data science, and domain knowledge to construct a bespoke solution over months or years. Instead of wrestling with scope creep, diverse teams are empowered to explore data how they each want to see it with rich comprehensive visualizations. Users can triangulate meaningful anomalies by applying machine learning algorithms to their key metrics, without requiring them to have a data science expertise.

In short, there is a simpler path to the digital intelligence needed improve performance, agility, and
digital experience of today's business.

Learn more about Corvil Intelligence Hub
Explore the Solution   Watch Now
Download the Solution Brochure
Learn about Self-Service Intelligence for Electronic Trading

David Murray

David Murray, Chief Marketing & Business Development Officer
Pico is a leading provider of technology services for the financial markets community. Pico provides a best-in-class portfolio of innovative, transparent, low-latency markets solutions coupled with an agile and expert service delivery model. Instant access to financial markets is provided via PicoNet™, a globally comprehensive network platform instrumented natively with Corvil analytics and telemetry. Clients choose Pico when they want the freedom to move fast and create an operational edge in the fast-paced world of financial markets.